On Intra Prediction for Screen Content Video Coding

Screen content coding (SCC) is becoming increasingly important in various applications, such as desktop sharing, video conferencing, and remote education. When compared to natural camera- captured content, screen content has different characteristics, in particular sharper edges. In this paper, we propose a novel intra prediction scheme for screen content video. In the proposed scheme, bilinear interpolation in angular intra prediction in HEVC is selectively replaced by nearest-neighbor intra prediction to preserve the sharp edges in screen content video. We present three different variants of the proposed nearest neighbor prediction algorithm: two implicit methods where both the encoder, and the decoder derive whether to perform nearest neighbor prediction or not based on either (a) the sum of the absolute difference, or (b) the difference between the boundary pixels from which prediction is performed; and another variant where Rate-Distortion-Optimization (RDO) search is performed at the encoder to decide whether or not to use the nearest neighbor interpolation, and explicitly signaled to the decoder. We also discuss the various underlying trade-offs in terms of the complexity of the three variants. All the three proposed variants provide significant gains over HEVC, and simulation results show that average gains of 3.3% BD-bitrate in Intra-frame coding are achieved by the RDO variant for screen content video. To the best of our knowledge, this is the first paper that 1) points out current HEVC intra prediction scheme with bilinear interpolation does not work efficiently for screen content video and 2) uses different filters adaptively in the HEVC intra prediction interpolation.